Occam

On apppeals to Occam, or whichever way one wants to spell it:

Popper mentions somewhere, but doesn't elaborate much, that when given
the choice between two equally falsifiable theories (whichever way one
defines that) one should prefers the simpler. This is as far as I can
see equivalent to Occam's advice. On how to make out which one is the
simpler, all he can suggest is something like counting the number of
symbols required to write the theory.

But that introduces a whole new type of uncertainty into the procedure:
In what language, whether natural or made up, should theories be
written? His offhand remark seems to suggest that there exists such a
universal and already canonical language in which theories must of
necessity be written. Surprise: there is no such language! Any of you
who has been prolonged contact with logicians and computer scientists
know that there are bitter wars fought over the correct symbolic
representations of states of affairs. Resorting to the deistic Anglo-
Saxon habit of writing Logic and Truth and Language with a capital
letter will not cause a Logic, a Truth and a Language to exist on the
other bank of the river that divides matter from the realm of ideas,
because: they don't exist! Logics, truths and languages may exist,
Logic, Truth and Language don't.

Therefore: Appealing to Occam is asking to be graded on style, until
someone defines and we all agree on a Universal Scientific Symbolic
Language. Having studied Artificial Intelligence and its various
proposed languages for some years I know whereof I speak.


Torsten